Solving Markov Decision Networks using Incremental Pruning